Visual C++ 6 Unleashed

By MICKEY WILLIAMS and David Bennett

• Table of Contents
  • Copyright
  • About the Authors
  • About the Contributors
  • Acknowledgments
  • Tell Us What You Think!
  • Introduction
  • How to Use This Book
  • What You Need to Use This Book
  • What's New in Visual C++ 6.0
  • Contacting the Main Author
  • Part I: Introduction
  • Chapter 1. The Visual C++ 6.0 Environment
  • Part II: MFC Programming
  • Chapter 2. MFC Class Library Overview
  • Chapter 3. MFC Message Handling Mechanism
  • Chapter 4. The Document View Architecture
  • Chapter 5. Creating and Using Dialog Boxes
  • Chapter 6. Working with Device Contexts and GDI Objects
  • Chapter 7. Creating and Using Property Sheets
  • Chapter 8. Working with the File System
  • Chapter 9. Using Serialization with File and Archive Objects
  • Part III: Internet Programming with MFC
  • Chapter 10. MFC and the Internet Server API (ISAPI)
  • Chapter 11. The WinInet API
  • Chapter 12. MFC HTML Support
  • Part IV: Advanced Programming Topics
  • Chapter 13. Using the Standard C++ Library
  • Chapter 14. Error Detection and Exception Handling Techniques
  • Chapter 15. Debugging and Profiling Strategies
    • MFC Support for Debugging
    • Techniques for Debugging Your MFC Application
    • Using the DevStudio IDE Debugger
    • Using Spy++
    • Using the OLE-COM Object Viewer
    • Using the Process Viewer
    • Performing Remote Debugging
    • Troubleshooting
    • Profiling Your Application
    • Summary
  • Chapter 16. Multithreading
  • Chapter 17. Using Scripting and Other Tools to Automate the Visual C++ IDE
  • Part V: Database Programming
  • Chapter 18. Creating Custom AppWizards
  • Chapter 19. Database Overview
  • Chapter 20. ODBC Programming
  • Chapter 21. MFC Database Classes
  • Chapter 22. Using OLE DB
  • Chapter 23. Programming with ADO
  • Part VI: MFC Support for COM and ActiveX
  • Chapter 24. Overview of COM and Active Technologies
  • Chapter 25. Active Documents
  • Chapter 26. Active Containers
  • Chapter 27. Active Servers
  • Chapter 28. ActiveX Controls
  • Part VII: Using the Active Template Library
  • Chapter 29. ATL Architecture
  • Chapter 30. Creating COM Objects Using ATL
  • Chapter 31. Creating ActiveX Controls Using ATL
  • Chapter 32. Using ATL to Create MTS and COM+ Components
  • Part VIII: Finishing Touches
  • Chapter 33. Adding Windows Help
  • Part IX: Appendix

Profiling Your Application

Profiling your application enables you to determine which areas of your code are working efficiently. To accomplish this, you must use the profiler. The profiler is an analysis tool used to examine the runtime behavior of your programs. This analysis tool produces feedback exposing sections of code that are taking a long time to execute or are not being executed at all.

Remember that profiling is a tuning process, not a one-shot deal. The intention of this tool is to make your programs run better. It is not meant to find bugs or to be used for debugging purposes.

The main reasons for using the profiler follow:

• To determine whether an algorithm is effective
• To determine whether a function is too few or too many times (with respect to the problem domain)
• To determine whether a piece of code is being covered by the software testing procedures

The first reason represents a timing issue. The second represents a counting issue, and the third represents a coverage issue.

You can use two types of profiling to analyze code:

• FunctionProfiling locates code that is inefficient. FunctionProfiling generally is faster, because there is usually less information to collect.
• LineProfiling checks how many times each line is executed, plus the ones that are not executed at all. You use LineProfiling to verify the validity of an algorithm.

Building the Project for Function Profiling

Before you use the profiler, you must build the project with profiling enabled. To enable profiling, do the following:

1. From the menu bar, choose Project, Settings.
2. Select the Link tab.
3. From the Category drop-down list box, select General.
4. Click the check mark in the Enable Profiling check box to enable the option.
5. Click OK.
6. Build the project.

Note that incremental linking is turned off when enabling the Enable Profiling check box. Clear the Enable Profiling option to reenable incremental linking.

Line Profiling

There are two types of line profiling:

• Line counting shows how many times each line was executed.
• Line coverage shows which lines were executed at least once.

Line profiling uses the debugging information from the .EXE file of the application to trigger the profiler.

To implement line profiling when building the project, follow these steps:

1. From the menu bar, choose Project, Settings.
2. Select the Link tab.
3. From the Category drop-down list box, select General.
4. Enable the Enable Profiling check box.
5. Click OK.
6. Build the project.
7. From the menu bar, choose Project, Settings.
8. Select the Link tab.
9. From the Category drop-down list box, select General.
10. Enable the Enable Profiling check box.
11. Enable the Generate Debug Info check box.
12. Select the C/C++ tab.
13. In the Category drop-down list box, select General.
14. In the Debug Info drop-down list box, select Program Database or Line Numbers Only.
15. Click OK.
16. Build the project.

After the build is complete, the project is ready to be profiled.

Profiling with PREP, PROFILE, and PLIST

Profiling consists of three separate programs:

• PREP
• PROFILE
• PLIST

The IDE runs all these programs for you after you choose the standard option from the Profile dialog box. The arguments are passed automatically from the PREP program.

Using batch files to run these programs enables you to get the most out of profiling. These files give you more options and control over profiling the applications. You can run the batch files from the command prompt or the Profile dialog box. PLIST output is routed to the output window when using the Profile dialog box. You can redirect the command-line output, such as to a file or to an output device.

The PREP program always is called twice, because there are two phases that are performed when profiling. The first phase is done before the actual profiling is done, and the second phase is done after the profiling.

To transfer information between profiling steps, three files are implemented: .PBI, .PBO, and .PBT.

.PBI is the file extension for the profiler batch input file. This file provides condensed information to the profiler. The first time the profiler is run on a program, the .PBI file is generated by the PREP program.

.PBO is the file extension for the profiler batch output file. This file is used as an intermediate file used by the profiler. Its function is to transfer information between profiling steps.

The .PBT (profiler) file is generated by the PREP program and is used as input to the PLIST program. With this file, the PLIST program generates a profile of the source code.

All these files are intermediate files used by the profiler to transfer information and data between profiling steps.

A profiler batch file might look like the following:

PREP /OM /FT /EXC test.lib %1
if errorlevel == 1 goto done
PROFILE %1 %2 %3 %4 %5 %6 %7 %8 %9
if errorlevel == 1 goto done
PREP /M %1
if errorlevel == 1 goto done
PLIST /SC %1 >%1.lst
:done

The batch file options will make more sense after you examine each of the programs.

PREP

The PREP program has two phases during a normal profiling operation. In phase 1, it reads an executable file and creates .PBT and .PBI files.

Phase 2 consists of reading .PBT and .PBO files, and then writing a new .PBT file for PLIST.

The syntax for PREP follows:

PREP [options][programname1][programname2...programnameN]

The command line is read from left to right. Table 15.2 provides a list of PREP options. Note that since the command is read from left to right, the rightmost options override the contradictory options to the left. None of the options are case-sensitive, and all must be prefixed by a forward slash or a dash. Options are separated by spaces.

Table 15.2. PREP Options

To specify individual library, object, and application source files, use the /INC and /EXC options.

To specify line numbers with source files, you would use syntax such as the following:

/EXCALL /INC example.cpp(12-40,50-80)

The following syntax specifies all line numbers from 80 on:

/EXCALL /INC example.cpp(80-0).

Only lines 12 to 40 and 50 to 80 are included from the file example.cpp.

If you need to specify all the lines in a module, you can specify the .OBJ file as the following:

/EXCALL /INC example.obj
/EXCALL /INC test.cpp(50-0)

PROFILE

The PROFILE program profiles an application and stores the results in a .PBO file. PROFILE is used after a .PBI file is created with PREP.

The syntax for PREP follows:

PROFILE [options] programname [programargs]

The command line is read from left to right. Table 15.3 provides a list of PROFILE options. Note that because the command is read from left to right, the rightmost options override the contradictory options to the left. None of the options are case sensitive, and all must be prefixed by a forward slash or a dash. Options are separated by spaces.

If a .PBO filename is not specified on the command line, the base name of the .PBI file with a .PBO extension is used. If neither a .PBI nor a .PBO file is specified, a program name with .PBI and .PBO extensions is used.

Table 15.3. PROFILE Options

On the PROFILE command line, the filename of the program that is to be profiled must be designated. If no extension is given, the PROFILE assumes the .EXE extension. The program filename can be followed with command-line arguments.

PLIST

The PLIST program converts the results from a .PBT file into a formatted text file.

The syntax for PLIST follows:

PLIST [options] inputfile

The command line is read from left to right. Table 15.4 provides a list of PLIST options. Note that because the command is read from left to right, the rightmost options override the contradictory options to the left. None of the options are case-sensitive, and all must be prefixed by a forward slash or a dash. Options are separated by spaces.

The results of PLIST are sent to STDOUT by default. You can use the greater-than (>) character to redirect the output to a file or device.

Table 15.4. PLIST Options

You must run PLIST in the same directory in which the profiled program was compiled.

If a PLIST variable is not specified, default values for PLIST are supplied. The default used depends on the profile type.

Table 15.5 shows the default values used when the PLIST environment variable is not specified.

Table 15.5. Default Values Used with PLIST

Share This